TY  - RPRT
TI  - Are Vision-Language-Action Models Robust to One-Step Observation Perturbations?
AU  - Shojiro Yamabe
AU  - Jun Sakuma
PY  - 2026
UR  - https://arxiv.org/abs/2609.32550
ID  - 2609.32550
ER  -